Challenges
Datasets
Workspaces
Papers
Discussions
Leaderboard
Log inSign up
Challenges
Datasets
Workspaces
Papers
Discussions
Leaderboard
Blog
Job Board
Q3AS

© 2026 Aqora Quantum S.A.S.

TermsPrivacyLegal Notice

arXiv:2408.05629 · Physical Review X

Quantum-Secure Multiparty Deep Learning

Kfir Sulimany, Sri Krishna Vadlamani, Ryan Hamerly, Prahlad Iyengar, Dirk Robert Englund

Published Oct 31, 2025• 13 citations
arXivHTMLPDF

Abstract

Secure multiparty computation enables the joint evaluation of multivariate functions across distributed users while ensuring the privacy of their local inputs. This field has become urgent due to the demand for computationally intensive deep learning inference. These computations are typically offloaded to cloud servers, leading to vulnerabilities. To solve this problem, we introduce a linear algebra engine that leverages the quantum nature of light for information-theoretically secure multiparty inference using telecommunication components. We apply this linear algebra engine to deep learning and derive rigorous upper bounds on the information leakage of both the deep neural network weights and the client’s data, enabling double-blind operations. Applied to the modified National Institute of Standards and Technology classification task, we obtain test accuracies exceeding 95% while guaranteeing leakage of less than 0.1 bits per weight and data element. This leakage is an order of magnitude below the minimum bit precision required for accurate deep learning using state-of-the-art quantization techniques. Our work lays the foundation for practical quantum-secure computation and unlocks secure cloud deep learning as a field.

DOI
10.1103/k8wg-qmbh

Related papers

Quantum-aided secure deep neural network inference on real quantum computers

Hanqiao Yu, Xuebin Ren, Cong Zhao +2 more

Deep neural networks (DNNs) are phenomenally successful machine learning methods broadly applied to many different disciplines. However, as complex two-party computations, DNN inference using classical cryptographic methods cannot achieve unconditional security, raising concern on security risks of DNNs' application to sensitive data in many domains. We overcome such a weakness by introducing a quantum-aided security approach. We build a quantum scheme for unconditionally secure DNN inference based on quantum oblivious transfer with an untrusted third party. Leveraging DNN's noise tolerance, our approach enables complex DNN inference on comparatively low-fidelity quantum systems with limited quantum capacity. We validated our method using various applications with a five-bit real quantum computer and a quantum simulator. Both theoretical analyses and experimental results demonstrate that our approach manages to operate on existing quantum computers and achieve unconditional security with a negligible accuracy loss. This may open up new possibilities of quantum security methods for deep learning.

Scientific Reports•Nov 5, 2023•8 citations

Quantum-Secure Deep Learning in Optical Networks

Kfir Sulimany, Sri Krishna Vadlamani, Ryan Hamerly +2 more

The demand for cloud-based deep learning has intensified the need for secure computation. Our quantum-secure deep learning protocol leverages quantum properties of light and telecommunication components for scalable, secure cloud-based deep learning in optical networks.

Optical Fiber Communication Conference•Jan 1, 2025•0 citations

Quantum-Secure Cloud Deep Learning in Optical Networks

Kfir Sulimany, Sri Krishna Vadlamani, Ryan Hamerly +2 more

The demand for cloud-based deep learning has intensified the need for secure computation. Our quantum-secure deep learning protocol leverages quantum properties of light and telecommunication components for scalable, secure cloud-based deep learning in optical networks.

Conference on Lasers and Electro-Optics•Jan 1, 2025•0 citations

Reverse Delegated Training and Private Inference via Perfectly-Secure Quantum Homomorphic Encryption

Sergio A. Ortega, Miguel A. Martin-Delgado

Quantum machine learning in cloud environments requires protecting sensitive data while enabling remote computation. Here we demonstrate the first realistic implementations of a perfectly-secure quantum homomorphic encryption (QHE) scheme applied to quantum neural networks (QNN). Using efficient Clifford+$T$ decomposition, we implement quantum convolutional neural networks for two complementary scenarios: (i) reverse delegated training, where encrypted data from multiple providers trains a user's network via federated aggregation; (ii) private inference, where users process encrypted data with remote quantum networks. Moreover, analysis of server circuit privacy reveals probabilistic model protection through Pauli gate concealment. These results establish perfectly-secure QHE as a practical framework for multi-party quantum machine learning.

quant-pharXiv:2602.12712•Feb 16, 2026•0 citations

Further Investigation on Classical Multiparty computation using Quantum Resources

Hao Cao, Wenping Ma, Ge Liu +1 more

The tremendous development of cloud computing and network technology makes it possible for multiple people with limited resources to complete a large-scale computing with the help of cloud servers. In order to protect the privacy of clients, secure multiparty computation plays an important role in the process of computing. Recently, Clementi et al[\textcolor[rgb]{0.00,0.07,1.00}{Phys. Rev. A {\bf 96}, 062317(2017)}] proposed a secure multiparty computation protocol using quantum resources. In their protocol, utilizing only linear classical computing and limited manipulation of quantum information, a method of computing $n-variable$ symmetric Boolean function $f(x_1, x_2, \cdots, x_n)$ with degree 2 is proposed, and all clients can jointly compute $f(x_1, x_2, \cdots, x_n)$ without revealing their private inputs with the help of a sever. They proposed an open problem: are there more simple nonlinear functions like the one presented by them that can be used as subroutines for larger computation protocols? We will give the answer to this question in this paper. Inspired by Clementi et al's work, we continue to explore the quantum realization of Boolean functions. First, we demonstrate a way to compute a class of $n-variable$ symmetric Boolean function $f_n^k$ by using single-particle quantum state $|0\rangle$ and single-particle unitary operations $U_k$. Second, we show that each $n-variable$ symmetric Boolean function can be represented by the linear combination of $f_n^k(k=0,1,\cdots,n)$ and each function $f_n^k(2\leq k\leq n)$ can be used to perform secure multiparty computation. Third, we propose an universal quantum implementation method for arbitrary $n-variable$ symmetric Boolean function $f(x_1, x_2, \cdots, x_n)$. Finally, we demonstrate our secure multiparty computation protocol on IBM quantum cloud platform.

arXiv:1812.06425•Dec 16, 2018•0 citations

Quantum Secure Multiparty Computation for Privacy-Preserving AI

Made Dwi Setyadhi Mustika

Quantum Secure Multiparty Computation (QMPC) is a new way to use cryptography that lets a lot of people work together to compute a function over their private inputs while keeping those inputs secret, even if attackers have quantum computing power. As AI becomes more common in fields like healthcare, banking, and self- driving cars, secure collaborative learning becomes more and more important. But the rise of quantum computers is a big threat to the encryption methods that are commonly used in modern multiparty computation protocols. QMPC is a very secure and scalable way to protect privacy in AI. It does this by combining quantum-safe oblivious transfer (QOT), quantum key distribution (QKD), and quantum oblivious linear evaluation (qOLE). This study shows a full QMPC architecture for AI systems that are both federated and decentralized. It lets you train and use private neural networks and is safe from quantum attacks because it uses safe aggregation and encrypted processing protocols. The architecture lets AI do a lot of things, like private set intersection, federated learning, and collaborative model training, without giving away user data. A thorough security analysis based on postquantum assumptions and the universal composability paradigm makes sure that the system can withstand attackers who are adaptive, colluding, or quantum-capable. When tested against datasets like MNIST and CIFAR-10, the protocol gets results that are as accurate as plaintext models with very little extra work on the computer or network. Experiments show that when the accuracy of the model drops by less than 1%, all data privacy is still protected. This study looks at how QMPC might help make AI safer in sensitive areas like medical diagnostics and financial forecasts, as well as how it might help companies follow data protection laws like GDPR and HIPAA. Auditability and consent procedures are two of the legal and moral issues that need to be thought about. The study also talks about problems with scalability and suggests an ideal communication plan and methods like circuit pruning, quantization, and hybrid classical-quantum communication models. We look at what QMPC could do in relation to secure federated analytics, real-time collaborative robotics, and processing encrypted smart contracts. In conclusion, QMPC for privacy-preserving AI is no longer just a theory; it is now a necessary, possible, and realistic way of doing things in a world that is quickly moving toward quantum supremacy. By using quantum-resistant encryption and distributed machine learning together, QMPC creates a way to get to safe, understandable, and decentralized AI that will work in the future. This study sets the stage for future progress at the crossroads of postquantum security and collaborative intelligence.

International Journal of Quantum Computing and Artificial Intelligence (IJQCAI)•Dec 24, 2025•0 citations
Secure multiparty computation enables the joint evaluation of multivariate functions across distributed users while ensuring the privacy of their local inputs. This field has become urgent due to the demand for computationally intensive deep learning inference. These computations are typically offloaded to cloud servers, leading to vulnerabilities. To solve this problem, we introduce a linear algebra engine that leverages the quantum nature of light for information-theoretically secure multiparty inference using telecommunication components. We apply this linear algebra engine to deep learning and derive rigorous upper bounds on the information leakage of both the deep neural network weights and the client’s data, enabling double-blind operations. Applied to the modified National Institute of Standards and Technology classification task, we obtain test accuracies exceeding 95% while guaranteeing leakage of less than 0.1 bits per weight and data element. This leakage is an order of magnitude below the minimum bit precision required for accurate deep learning using state-of-the-art quantization techniques. Our work lays the foundation for practical quantum-secure computation and unlocks secure cloud deep learning as a field.
Deep neural networks (DNNs) are phenomenally successful machine learning methods broadly applied to many different disciplines. However, as complex two-party computations, DNN inference using classical cryptographic methods cannot achieve unconditional security, raising concern on security risks of DNNs' application to sensitive data in many domains. We overcome such a weakness by introducing a quantum-aided security approach. We build a quantum scheme for unconditionally secure DNN inference based on quantum oblivious transfer with an untrusted third party. Leveraging DNN's noise tolerance, our approach enables complex DNN inference on comparatively low-fidelity quantum systems with limited quantum capacity. We validated our method using various applications with a five-bit real quantum computer and a quantum simulator. Both theoretical analyses and experimental results demonstrate that our approach manages to operate on existing quantum computers and achieve unconditional security with a negligible accuracy loss. This may open up new possibilities of quantum security methods for deep learning.
Quantum machine learning in cloud environments requires protecting sensitive data while enabling remote computation. Here we demonstrate the first realistic implementations of a perfectly-secure quantum homomorphic encryption (QHE) scheme applied to quantum neural networks (QNN). Using efficient Clifford+TTT decomposition, we implement quantum convolutional neural networks for two complementary scenarios: (i) reverse delegated training, where encrypted data from multiple providers trains a user's network via federated aggregation; (ii) private inference, where users process encrypted data with remote quantum networks. Moreover, analysis of server circuit privacy reveals probabilistic model protection through Pauli gate concealment. These results establish perfectly-secure QHE as a practical framework for multi-party quantum machine learning.
The tremendous development of cloud computing and network technology makes it possible for multiple people with limited resources to complete a large-scale computing with the help of cloud servers. In order to protect the privacy of clients, secure multiparty computation plays an important role in the process of computing. Recently, Clementi et al[\textcolor[rgb]{0.00,0.07,1.00}{Phys. Rev. A {\bf 96}, 062317(2017)}] proposed a secure multiparty computation protocol using quantum resources. In their protocol, utilizing only linear classical computing and limited manipulation of quantum information, a method of computing n−variablen-variablen−variable symmetric Boolean function f(x1,x2,⋯ ,xn)f(x_1, x_2, \cdots, x_n)f(x1​,x2​,⋯,xn​) with degree 2 is proposed, and all clients can jointly compute f(x1,x2,⋯ ,xn)f(x_1, x_2, \cdots, x_n)f(x1​,x2​,⋯,xn​) without revealing their private inputs with the help of a sever. They proposed an open problem: are there more simple nonlinear functions like the one presented by them that can be used as subroutines for larger computation protocols? We will give the answer to this question in this paper. Inspired by Clementi et al's work, we continue to explore the quantum realization of Boolean functions. First, we demonstrate a way to compute a class of n−variablen-variablen−variable symmetric Boolean function fnkf_n^kfnk​ by using single-particle quantum state ∣0⟩|0\rangle∣0⟩ and single-particle unitary operations UkU_kUk​. Second, we show that each n−variablen-variablen−variable symmetric Boolean function can be represented by the linear combination of fnk(k=0,1,⋯ ,n)f_n^k(k=0,1,\cdots,n)fnk​(k=0,1,⋯,n) and each function fnk(2≤k≤n)f_n^k(2\leq k\leq n)fnk​(2≤k≤n) can be used to perform secure multiparty computation. Third, we propose an universal quantum implementation method for arbitrary n−variablen-variablen−variable symmetric Boolean function f(x1,x2,⋯ ,xn)f(x_1, x_2, \cdots, x_n)f(x1​,x2​,⋯,xn​). Finally, we demonstrate our secure multiparty computation protocol on IBM quantum cloud platform.
Quantum Secure Multiparty Computation (QMPC) is a new way to use cryptography that lets a lot of people work together to compute a function over their private inputs while keeping those inputs secret, even if attackers have quantum computing power. As AI becomes more common in fields like healthcare, banking, and self- driving cars, secure collaborative learning becomes more and more important. But the rise of quantum computers is a big threat to the encryption methods that are commonly used in modern multiparty computation protocols. QMPC is a very secure and scalable way to protect privacy in AI. It does this by combining quantum-safe oblivious transfer (QOT), quantum key distribution (QKD), and quantum oblivious linear evaluation (qOLE). This study shows a full QMPC architecture for AI systems that are both federated and decentralized. It lets you train and use private neural networks and is safe from quantum attacks because it uses safe aggregation and encrypted processing protocols. The architecture lets AI do a lot of things, like private set intersection, federated learning, and collaborative model training, without giving away user data. A thorough security analysis based on postquantum assumptions and the universal composability paradigm makes sure that the system can withstand attackers who are adaptive, colluding, or quantum-capable. When tested against datasets like MNIST and CIFAR-10, the protocol gets results that are as accurate as plaintext models with very little extra work on the computer or network. Experiments show that when the accuracy of the model drops by less than 1%, all data privacy is still protected. This study looks at how QMPC might help make AI safer in sensitive areas like medical diagnostics and financial forecasts, as well as how it might help companies follow data protection laws like GDPR and HIPAA. Auditability and consent procedures are two of the legal and moral issues that need to be thought about. The study also talks about problems with scalability and suggests an ideal communication plan and methods like circuit pruning, quantization, and hybrid classical-quantum communication models. We look at what QMPC could do in relation to secure federated analytics, real-time collaborative robotics, and processing encrypted smart contracts. In conclusion, QMPC for privacy-preserving AI is no longer just a theory; it is now a necessary, possible, and realistic way of doing things in a world that is quickly moving toward quantum supremacy. By using quantum-resistant encryption and distributed machine learning together, QMPC creates a way to get to safe, understandable, and decentralized AI that will work in the future. This study sets the stage for future progress at the crossroads of postquantum security and collaborative intelligence.